Fiber Data Hub DSI Studio ↗ ☰
Featured datasets › Agent Guide▾
Fiber Data Hub › Featured datasets

Fiber Data Hub Agent Guide

This repository serves https://brain.labsolver.org, the web portal for the Fiber Data Hub.

AI agents may access Fiber Data Hub data through either of two supported routes:

  1. Direct web/GitHub access
  2. DSI Studio Fiber Data Hub commands

Both routes are valid. Choose the route that best fits the task and the tools currently available. Do not require DSI Studio when ordinary web/GitHub access is sufficient, and do not reimplement DSI Studio analysis when DSI Studio is available and already provides the operation.

Route 1: Direct web/GitHub access

Use this route for dataset discovery, release metadata, QC lookup, file listing, downloading public files, or external Python workflows.

Entry points

Public Fiber Data Hub data are distributed primarily as GitHub release assets in repositories under organizations including:

The current repository list is maintained on https://brain.labsolver.org. Use the portal rather than assuming the list above is exhaustive.

Discover repositories and datasets

A dataset is normally addressed as:

owner/repository/tag

Examples:

data-hcp/lifespan/hcp-ya
data-openneuro/brain/ds004299

GitHub API examples:

https://api.github.com/users/{owner}/repos
https://api.github.com/repos/{owner}/{repository}/releases
https://api.github.com/repos/{owner}/{repository}/releases/tags/{tag}

The release-tag response contains an assets array. Use each asset’s name, size, and browser_download_url to inspect or download files.

Search QC data

Centralized QC tables are release assets at:

https://api.github.com/repos/frankyeh/FiberDataHub/releases/tags/qc-data

Select TSV assets matching the requested owner/repository/dataset, download them using browser_download_url, and search the table contents. The Fiber Data Hub homepage contains a current Python example.

File formats

Common Fiber Data Hub files include:

For direct programmatic reading or conversion of DSI Studio formats, follow:

https://dsi-studio.labsolver.org/doc/cli_data.html

Use the documented format rather than guessing the binary layout.

Route 2: DSI Studio Fiber Data Hub commands

Use this route when DSI Studio is available and the task benefits from its built-in Hub integration, local GUI/session context, opening downloaded data, visualization, tractography, or subsequent DSI Studio analysis. It may also be useful when direct public access is insufficient for the requested dataset.

Current Fiber Data Hub commands are:

hub_repo
hub_tags <repo>
hub_files <repo> [tag] [text] [offset] [limit]
hub_open <repo> <tag> <file>
hub_show <repo> <tag> [file]
hub_download <repo> [tag] <file> <dir>

Here <repo> is normally owner/repository.

For hub_files, empty tag/text values mean match all; tag and text may be regular expressions. For hub_download, <file> is a wildcard pattern such as *.qsdr.fz and may match multiple files.

When using DSI Studio through an AI-agent connection, invoke these commands through the DSI Studio command interface provided by that session. Do not substitute web scraping if the requested operation specifically requires opening or analyzing the data inside DSI Studio.

Choosing a route

Choose based on the requested result rather than a fixed preference:

If one route fails or is unavailable, try the other when it can satisfy the same request.

Data-use rules

Fiber Data Hub derivatives originate from multiple source datasets with different licenses and data-use requirements.

Ground truth

For current dataset availability, metadata, download links, licensing notes, and citations, use the live Fiber Data Hub pages and corresponding GitHub releases. Do not rely on a hard-coded dataset count or an old cached repository list when live information is available.